Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Auxiliary problem on different domain
Posted 21 mag 2014, 07:17 GMT-4 Fluid & Heat, Geometry, Studies & Solvers Version 4.4 1 Reply
Please login with a confirmed email address before reporting spam
Good afternoon!
I am simulating the reaction-diffusion problem in 2 domains separated by a membrane with transmission conditions on this membrane.
I have 3 reaction-diffusion equations for 3 functions: u, u2, u3.
u (t, x, y) and u2 (t, x, y):
- are defined on 2 2d domains (which are squares actually) with the adjacent boundary (= membrane = S = (x, 0))
- on the S each has Neumann condition defined by the integral of u3
u3 (t, x, y, z):
- is defined on 3d domain ([0, 1] x [0, 1] x [-1, 1])
- has Dirichlet boundary condition on 2 (out of 6) boundaries defined by the functions u and u2 (u3(t, x, y, 1) = u (t, x, 0); u3(t, x, y, -1) = u2(t, x, 0))
I define 2 different components: one for 2d problems, another - for auxiliary 3d problem.
I receive the following problem: Failed to evaluate variable Jacobian. Under variable u and u2 are understood in the part where they enter boundary conditions for u3.
I think that maybe the problem is related to the fact that the functions are defined on different domains and in different components, but I do not know the way to define the equations in one component.
Could you give me a piece of advice how to overcome this error?
Another problem which I have is the abortion when I try to switch to the tab with the second (3d) geometry. It is displayed on the other tabs (e.g. on the Convection-Diffusion Equation tab), but when I try to switch to the Geometry Comsol shuts down.
I would appreciate any help you could provide.
Thank you in advance!
I am simulating the reaction-diffusion problem in 2 domains separated by a membrane with transmission conditions on this membrane.
I have 3 reaction-diffusion equations for 3 functions: u, u2, u3.
u (t, x, y) and u2 (t, x, y):
- are defined on 2 2d domains (which are squares actually) with the adjacent boundary (= membrane = S = (x, 0))
- on the S each has Neumann condition defined by the integral of u3
u3 (t, x, y, z):
- is defined on 3d domain ([0, 1] x [0, 1] x [-1, 1])
- has Dirichlet boundary condition on 2 (out of 6) boundaries defined by the functions u and u2 (u3(t, x, y, 1) = u (t, x, 0); u3(t, x, y, -1) = u2(t, x, 0))
I define 2 different components: one for 2d problems, another - for auxiliary 3d problem.
I receive the following problem: Failed to evaluate variable Jacobian. Under variable u and u2 are understood in the part where they enter boundary conditions for u3.
I think that maybe the problem is related to the fact that the functions are defined on different domains and in different components, but I do not know the way to define the equations in one component.
Could you give me a piece of advice how to overcome this error?
Another problem which I have is the abortion when I try to switch to the tab with the second (3d) geometry. It is displayed on the other tabs (e.g. on the Convection-Diffusion Equation tab), but when I try to switch to the Geometry Comsol shuts down.
I would appreciate any help you could provide.
Thank you in advance!
Attachments:
1 Reply Last Post 26 mag 2014, 04:32 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Dear Anastasia,
I found some issues in your model that need to be revised. The issues emerge partly from handling problems; therefor I suggest that you take a look at the model library with many different examples and step-by-step instructions.
Let's start with your first component:
I am unsure if u and u2 is the same magnitude. If that is the case I would rather delete the second application mode in comp1 and in case of convergence problems use the predefined smoothed step function for the initial value. It seems that you tried to re-implement this function under Global Definitions>Analytic. If you right-click Global Definitions you will see Functions>Step which would be the recommended way to model this.
The functional "integrate" is not defined. You can find the integration operators by right-clicking the Definitions-node in Component 1 and adding Component Couplings>Integration.
When coupling between different dimensions and components you have to make the variable globally known to all components and make sure that the values are mapped in the right way (which 2D-domain is mapped where as a surface in 3D) - the program cannot decide this for you ;-)
Coupling entities with the same dimension between model components of different dimension (here: from 3D-edge to 2D boundary, from 2D-domain to 3D-surface) can be defined via Component Couplings>Linear Extrusion. Instead of using u2, you will then use e.g. comp1.linext2(u2) in your dialog.
Please use also the context-sensitive Dynamic Help (the question mark at the upper right corner). It will describe in detail the functionality of each dialog window and its options.
In general it is a good idea to simplify the whole task by breaking up the different components and checking them separately. I suggest that you first fix the 2D-part. Just disable the 3D-part in the solver-node and check the corresponding error messages for component 1.
Hope that helps - good luck!
I found some issues in your model that need to be revised. The issues emerge partly from handling problems; therefor I suggest that you take a look at the model library with many different examples and step-by-step instructions.
Let's start with your first component:
I am unsure if u and u2 is the same magnitude. If that is the case I would rather delete the second application mode in comp1 and in case of convergence problems use the predefined smoothed step function for the initial value. It seems that you tried to re-implement this function under Global Definitions>Analytic. If you right-click Global Definitions you will see Functions>Step which would be the recommended way to model this.
The functional "integrate" is not defined. You can find the integration operators by right-clicking the Definitions-node in Component 1 and adding Component Couplings>Integration.
When coupling between different dimensions and components you have to make the variable globally known to all components and make sure that the values are mapped in the right way (which 2D-domain is mapped where as a surface in 3D) - the program cannot decide this for you ;-)
Coupling entities with the same dimension between model components of different dimension (here: from 3D-edge to 2D boundary, from 2D-domain to 3D-surface) can be defined via Component Couplings>Linear Extrusion. Instead of using u2, you will then use e.g. comp1.linext2(u2) in your dialog.
Please use also the context-sensitive Dynamic Help (the question mark at the upper right corner). It will describe in detail the functionality of each dialog window and its options.
In general it is a good idea to simplify the whole task by breaking up the different components and checking them separately. I suggest that you first fix the 2D-part. Just disable the 3D-part in the solver-node and check the corresponding error messages for component 1.
Hope that helps - good luck!